Accelerating the magic cycle of research breakthroughs and real-world applications
From earth science to genomics to quantum, we share the latest scientific breakthroughs from Google Research and how today’s powerful AI tools and platforms are accelerating innovation.
Last week at our flagship Research@ event in Mountain View, we shared some of Google Research’s latest announcements, from understanding earth to advancements in genomics to advancements in quantum computing. Our teams accelerate real-world solutions for products, businesses, science, and society by collaborating with colleagues across the organization. We find new research opportunities as research becomes real, accelerating innovation further. I refer to this potent, cyclical connection between research and impact in the real world as the “magic cycle of research.” These days, more powerful models, new agentic tools that help speed up scientific discovery, and open platforms and tools are accelerating this cycle significantly. This momentum is evident across all domains. Our most recent breakthroughs in research Google Earth AI, DeepSomatic, and Quantum Echoes were three of our most recent innovations that we presented at Research@MTV last week. Unprecedented planetary understanding with Google Earth AI Earth AI is a potent collection of geospatial AI models and reasoning intended to address significant global issues. It provides users with an unprecedented level of comprehension regarding global events. We have been creating cutting-edge geospatial AI models for floods, wildfires, cyclones, pollen, weather nowcasting and long-range forecasting, agriculture, population dynamics, AlphaEarth Foundations, and mobility for a number of years. We continue to advance, and these models, developed by teams across Google, are already assisting millions of people worldwide. We have just expanded access to our new Remote Sensing Foundations and new global Population Dynamics Foundations. And we can now share that our riverine flood models — expanded over the years to cover 700 million people in 100 countries — now provide forecasts covering over 2B people in 150 countries for significant riverine flood events.
Earth AI is a Google program that builds on our previous efforts. These enormous amounts of real-world imagery, population, and environmental data are integrated and synthesized in the most recent Earth AI updates that we have developed. The Earth AI geospatial reasoning agent is able to comprehend complex ideas and locate correlations across a variety of datasets and models by making use of LLMs and the reasoning capabilities that come with them. Earth AI capabilities are accessible even to non-experts thanks to this agent’s ability to answer complex questions in plain English. Users can quickly generate insights from business logic use cases and supply chain management to crisis resilience and international policy.
In our evaluations, Geospatial Reasoning Agent improved responses over baseline models that did not have access to Earth AI models and tools. In our technical report and research blog, we present the findings. Our Earth AI imagery models will soon power Google Earth’s Gemini capabilities, allowing users to search satellite imagery for objects. Plus, our powerful models are now available to trusted testers on Google Cloud. And we continue to hear from our partners about diverse important use cases, including testimonials from Give Directly, McGill and Partners, Cooper/Smith, WPP, WHO AFRO, Planet Labs and Airbus.
DeepSomatic and Cell2Sentence: A step toward cancer-fighting precision medicine Our most recent AI tool, DeepSomatic, was published in Nature Biotechnology. It is one of many designed to assist health professionals and scientists. DeepSomatic builds on Google’s ten years of genomics research. Since 2015, we’ve been building models like DeepConsensus and DeepVariant to help us better understand the genome. With these models, we’ve helped map human and non-human genomes and used this information to inform our understanding of disease.
Because of their intricate genetic signatures, some cancers may be targets for individualized treatments based on their particular mutations. So, we asked ourselves if we could sequence the genomes of these cancerous cells more precisely. Our new AI-powered open-source tool, DeepSomatic, was created as a result and is intended to assist medical professionals in comprehending genetic variants found in cancer cells. The model works by first turning genetic sequencing data into a set of images and then using a convolutional neural network to differentiate between the reference genome, the non-cancer germline variants in that individual, and the cancer-caused somatic variants in the tumor.
It is possible that identifying cancer variants will result in the development of brand-new treatments, and it may also assist physicians in deciding between treatments such as immunotherapy and chemotherapy. It is being used by our partners at Children’s Mercy to determine how and why a particular type of cancer affects a patient in order to develop individual cures. DeepSomatic follows other breakthroughs which share the same goal of using AI to help fight cancer. We also just released a 27 billion parameter foundation model for single-cell analysis, C2S-Scale, in collaboration with Google DeepMind. This builds upon our work from earlier this year, in collaboration with Yale, and recently generated a novel hypothesis about cancer cellular behavior. This may reveal a promising new path for the development of cancer therapies with additional clinical tests.
Quantum Echoes: A big step toward real-world applications
We are looking to our strategic, long-term investment in quantum computing to accelerate the subsequent exponential wave of scientific discovery. Our hardware milestone on the Willow chip, which will occur in late 2024, is built on decades of research. This work is supported by Michel Devoret, our Chief Scientist of Quantum Hardware, who together with with former Quantum AI hardware lead John Martinis, and John Clarke of the University of California, Berkeley, became 2025 Physics Nobel Laureates for their research in the 1980s that laid the groundwork for today’s superconducting qubits.
Now we’ve announced a new verifiable quantum advantage, published in the cover of Nature. Our “Quantum Echoes” algorithm runs on our Willow chip 13,000 times faster than the best classical algorithm on one of the world’s fastest supercomputers. It offers a new way to explain interactions between atoms in a real world molecule observed using nuclear magnetic resonance spectroscopy. This is the world’s first algorithm to demonstrate verifiable quantum advantage and points towards practical applications of quantum computing that are beyond the capabilities of classical computers.
Quantum computing has the potential to significantly advance drug design and contribute to the realization of fusion energy. Additionally, we are hopeful that we will begin to observe applications in the real world within five years given our most recent breakthrough.
